Self-powered microscale pumps based on analyte-initiated depolymerization reactions

Article abstract of DOI:10.1002/anie.201107787

Pump it up: Insoluble polymer films that depolymerize to release soluble monomeric products when exposed to a specific analyte act as a microscale pump. Products formed as a result of depolymerization amplify the signal and create a concentration gradient

Full text of DOI:10.1002/anie.201107787

.
Angewandte
Communications
DOI: 10.1002/anie.201107787
Microfluidic Pumps
Self-Powered Microscale Pumps Based on Analyte-Initiated
Depolymerization Reactions**
Hua Zhang, Kimy Yeung, Jessica S. Robbins, Ryan A. Pavlick, Meng Wu, Ran Liu,
Ayusman Sen,* and Scott T. Phillips*
Microscale pumps that are simultaneously capable of turning
“on” in response to specific analytes in solution and of
providing precise control over flow rate will be needed for
certain types of smart micro- and nanoscale devices. Access to
these types of pumps, however, has remained limited.[1] Two
general approaches for designing microscale pumps have
been proposed:[1–3] 1) those that are turned on and off by
input from the user; and 2) single-use pumps that are turned
on autonomously by exposure to a specific chemical signal,
and therefore do not need to be turned off. The ideal pump in
this second category would combine sensing and pumping
capabilities, and thus would enable pumping that is controlled
by the presence and concentration of a specific analyte.
Herein we describe the first examples of this second type of
microscale pump. We characterize this type of pump, and
establish both the chemical and physical–chemical foundation
upon which future applied efforts will be based.
The pumps consist of insoluble polymer films that
depolymerize to release soluble monomeric products when
exposed to a specific analyte (Figure 1).[4,5] Products formed
Figure 1. Design of self-powered micrometer-scale pumps that are
based on analyte-induced depolymerization of individual polymers
within a polymer film. Films were made from either tert-butyldimethyl-
silyl (TBS) end-capped poly(phthalaldehyde) (TBS-PPHA), which depo-
lymerizes in response to fluoride,[5] or poly(ethyl cyanoacrylate)
(PECA), which responds to base.[8]
as a result of the depolymerization reaction amplify the
signal[5] and create a concentration gradient that pumps fluids
(and insoluble particles) away from the bulk polymer owing
to a diffusiophoretic mechanism. These pumps are self-
powered, are capable of turning “on” in response to specific
analytes, and can be tuned to respond to a variety of analytes
ranging from small molecules to enzymes. They also provide
pumping velocities ranging from 0.1 mmsꢀ1 at low concen-
trations of analyte up to 11 mmsꢀ1 at high concentrations of
analyte. Moreover, the pumps are capable of moving fluids
and insoluble micrometer-scale particles directionally despite
viscous drag in low Reynolds number regions,[6] and despite
Brownian randomizations.[3a,7]
To demonstrate these types of pumps, we prepared
micrometer-scale polymer films on glass slides. The films
were composed of a polymer that is capable of depolymeriz-
ing when it responds to a specific chemical signal. We tested
two types of polymers: tert-butyldimethylsilyl (TBS) end-
capped poly(phthalaldehyde)[9,10] (TBS-PPHA),[5] which
depolymerizes in response to fluoride, and commercially
available poly(ethyl cyanoacrylate) (PECA),[8] which depo-
lymerizes and degrades in response to base (Figure 1).
The type of polymer used for the pumps is important: it
affects both the speed and magnitude of the pumping
response. For example, typical degradable polymers require
a separate reaction with the analyte for release of each
repeating unit. TBS-PPHA, in contrast, is a polymer that
depolymerizes rapidly (that is, seconds), continuously, and
completely from one end of the polymer to the other when the
TBS end-cap is cleaved from the polymer by reaction with
fluoride.[5] Thus, TBS-PPHA requires only a single equivalent
of fluoride to release 133 monomers (the TBS-PPHA
polymer used in this study contained about 133 repeating
units, although longer polymers could be used), meaning that
a single detection event causes a dramatic increase in the
concentration (and therefore gradient) of products in solu-
[*] H. Zhang, K. Yeung, J. S. Robbins, R. A. Pavlick, M. Wu, R. Liu,
Prof. A. Sen, Prof. S. T. Phillips
Department of Chemistry, The Pennsylvania State University
University Park, PA 16802 (USA)
E-mail: sphillips@psu.edu
asen@psu.edu
Homepage: http://research.chem.psu.edu/stpgroup/
http://research.chem.psu.edu/axsgroup/
[**] This work was supported by the NSF (DMR-0820404, CBET-
1014673), DARPA (N66001-09-1-2111), the Arnold and Mabel
Beckman Foundation, the Camille and Henry Dreyfus Foundation,
3M, Louis Martarano, and the Pennsylvania State University. We
appreciate the assistance of Prof. Darrell Velegol (Department of
Chemical Engineering) in calculating the power of the pumps.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201107787.
2400
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 2400 –2404

Angewandte
Chemie
tion. Poly(alkyl cyanoacrylates), in comparison, are a family
of biocompatible and biodegradable polymers that are non-
toxic and non-immunogenic to the human body.[11] The
mechanism for base-initiated depolymerization/degradation
of the PECA polymer backbone proceeds through several
pathways, including depolymerization from one end of the
polymer to the other, hydrolysis of the side-chain esters, as
well as other, less pronounced degradation pathways.[8,12]
Exposure of these polymer films either to fluoride (TBS-
PPHA) or base (PECA) causes depolymerization of the
polymers on the surface of the film, thus releasing monomeric
products into solution and initiating movement of the water
positioned above the film towards the film where there is
a high concentration of solute (that is, depoymerization
products; Figure 1). Water that moves initially in the z dir-
ection towards the film then moves away from the film
radially at the surface of the film, thus pumping away
exclusion was observed for the PECA pump. Moreover, the
direction of movement of the tracer particles in these types of
pumps is independent of the charge of the tracer particle (that
is, both negatively charged carboxylate functionalized tracers,
as well as positively charged amidine-functionalized tracers,
behave similarly; Supporting Information, Movie S3), thus
ruling out a pumping mechanism based upon an ionic
gradient.
Figure 3a,b further show that the pumping velocity
created by these pumps depends on the concentration of the
analyte. For example, Figure 3a shows the distribution of
particle velocities as a function of the concentration of
fluoride that was exposed to a TBS-PPHA pump. The particle
velocities in Figure 3a were measured at distances of 100 mm
Figure 2. Snapshots of exclusion zones around the TBS-PPHA polymer
film caused by analyte-induced radial pumping. The snapshots were
taken from Movie S1 in the Supporting Information. The dots in the
images are 6 mm diameter polystyrene tracer particles that reveal the
movement of the fluid. a) Image taken 0 s after exposure to 0.1m NaF
in distilled water at 258C, and b) 1000 s after exposure. The dotted line
in b) indicates the approximate boundary of the exclusion zone.
Movie S1 shows the entire pumping process.
insoluble particles (Figure 2; Supporting Information, Mov-
ies S1 and S2).
Figure 2 and Movie S1 show one example of this phenom-
enon. The images in Figure 2 show a film of TBS-PPHA that
is immersed in an aqueous solution (258C) that contains 6 mm
diameter polystyrene tracer particles and 0.1m NaF. No
movement of fluid is observed upon initial exposure to the
NaF solution (Figure 2a), but within seconds the tracer
particles begin to move radially away from the polymer film
to create an exclusion zone around the film (Figure 2b). In
the absence of fluoride, the particles do not move (TBS-
PPHA does not depolymerize),[5] and in the presence of 0.1m
NaCl, the average particle velocity is negligible (reaching only
0.1 mmsꢀ1), thus verifying the selectivity of the TBS-PPHA
pump for fluoride.[5] When exposed to 0.1m NaF, the TBS-
PPHA pumps are capable of continuous pumping for
> 15 min (settling of the 6 mm diameter polystyrene tracer
particles onto the surface of the glass slide precluded
measurements of longer pumping times).
Figure 3. Velocity (v) distributions of tracers. a) Exposure of TBS-PPHA
pumps to different fluoride concentrations. b) Exposure of PECA
pumps to aqueous solutions with different hydroxide concentrations.
c) At different distances from a TBS-PPHA pump when the pump was
exposed to 0.1m NaF at 258C. The particle velocities in (a) and (b)
were measured in the region 100 mm to 200 mm from the pump and
from 0 s to 7 s after exposure of the pump to the solution. The particle
velocities in (c) were measured over the time period of 5 s to 35 s after
exposure to fluoride. The relative frequency of particles is frel.
PECA pumps provide a similar exclusion zone as was
observed with the TBS-PPHA pumps, but in this case the
pumping response is caused by an aqueous solution of 0.01m
NaOH (Supporting Information, Movie S2). In a control
experiment using 0.01m NaCl instead of NaOH, no zone of
Angew. Chem. Int. Ed. 2012, 51, 2400 –2404
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2401

.
Angewandte
Communications
to 200 mm from the pump and at time points from 0 s to 7 s
after exposure to fluoride. There is a clear Gaussian
distribution of particle velocities at each concentration of
fluoride over this range of distances and time, yet the average
velocities of the particles increased as a function of the
concentration of fluoride, reaching an estimated maximum
velocity of 1.15 mmsꢀ1 at 0.1m NaF.
The effect of analyte concentration on pumping velocity
was observed with PECA pumps as well (Figure 3b), which
reached a maximum average pumping velocity of approx-
imately 11 mmsꢀ1 in a pH 14 aqueous solution (1m OHꢀ). This
velocity is approximately twice as fast as the average
swimming velocity of Escherichia coli (E. coli) (5–7 mmsꢀ1 at
surfaces[13]).
The pumping velocity for this type of pump, as with all
pumps, depends on the distance from the pump where the
velocity is measured. Figure 3c shows the change in average
particle velocity as a function of distance for a TBS-PPHA
pump over the time period of 5 s to 35 s after exposure to 0.1m
NaF at 258C. The data in Figure 3c reveals that the TBS-
PPHA pump is capable of providing average pumping
velocities of at least 0.8 ꢁ 0.4 mmsꢀ1 at distances over 1 mm
from the pump.
Figure 4. Proposed diffusiophoretic mechanism for enhanced rates of
diffusion of 2 mm PECA particles (maroon semi-spherical object) in
basic aqueous solutions. Note: As the particles degrade to some
extent over their entire surface creating multiple product sources, the
generated product gradient across the particle is not likely to be as
steep as that shown here in the idealized picture.
Our hypothesized mechanism for these pumps is depicted
in Figure 1. This mechanism is based on a non-electrolyte
diffusiophoretic model in which the concentration gradient of
monomeric products above the polymer film leads to move-
ment of water in the z direction from above the film towards
the film. This movement of water creates convective flow
within the confines of the test container that leads to the
movement of water initially in the z direction towards the
pump, and then radially away from the pump in the x,y plane.
To test this mechanism, we prepared 2 mm diameter
PECA particles (Supporting Information, Figure S1) as
a control system to evaluate whether analyte-induced depo-
lymerization would cause enhanced diffusion rates of the
PECA particles. Techniques for probing the mechanisms that
lead to motion of particles are well established,[14] therefore
we reasoned that control studies using PECA particles, rather
than films, might provide further insight into the mechanism
of the pumps.
creating a pressure gradient along the particle. This pressure
gradient drives fluid flow in the direction of higher product
concentration, thus pushing the particle in the opposite
direction. The depolymerization process is related to the
particle velocity (U) by the equation:[15,16]
kT
h
ð1Þ
U ¼
KLrC
where k is the Boltzmann constant, T is temperature, h is
viscosity, K is the Gibbs absorption length, L is the length of
the particle–product interaction, and C is the product
concentration. The observed enhanced diffusion coefficient
of the PECA particles in the presence of base can be related
to the ballistic velocity through the following equation:
U 2
4Dr
*
D ꢀD ¼
ð2Þ
The 2 mm-diameter PECA particles show enhanced
diffusivity (see Supporting Information) in basic solutions,
which we attribute to the depolymerization reaction and not
to the decrease in particle size that results over the course of
the depolymerization reaction. As depolymerization is an
endothermic, entropy-driven reaction, the enhanced diffusion
is not because of local heating. Instead, we propose that the
enhanced diffusivity arises from shape asymmetry and surface
imperfections in the particles that leads to greater exposure of
some areas on the particle to the analyte in solution
(Figure 4).
where D*ꢀD is the net enhancement in diffusion coefficient,
U is the particle velocity, and Dr is the rotational diffusion
coefficient (for a 2 mm particle Dr is calculated to be
0.23 sꢀ1).[14,17] Using the relationship described by Equa-
tion (2), the calculated velocity of the particles are
0.18 mmsꢀ1 in pH 8 solution and 0.25 mmsꢀ1 in pH 8.5
solution. These velocities compare well with the observed
pumping velocity for PECA pumps at pH 8 (0.21 ꢁ
0.12 mmsꢀ1). The agreement between the computed particle
velocity and the experimental pump velocities lead us to
conclude that the same mechanism probably causes both the
convective flow of water around the pumps and the enhanced
rate of diffusion of the PECA particles.
A larger area of exposure should lead to an increase in the
number of polymers that depolymerize, causing an uneven
increase in the local concentration of monomeric products
around the particle and an increased diffusivity owing to
a
non-electrolyte diffusiophoretic mechanism.[15,16] This
With an eye towards future applications, we broadened
the scope of these new types of autonomous pumps by
demonstrating the ability of TBS-PPHA pumps to move
mechanism arises from steric exclusion, which is generated
from the products interacting with the particle surface, thus
2402 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 2400 –2404

Angewandte
Chemie
enzyme b-d-glucuronidase (which is a specific marker of
E. coli)[18] by releasing two equivalents of fluoride per
enzymatic reaction (Supporting Information, Figure S2).
The released fluoride then turns “on” the pump. This
system differs from the experiments in Figure 2 and 3 in
that the small-molecule reagent (for example, 1) and the
pump must work in concert to provide selective pumping
when the enzyme is present. In theory, reagent 1 could be
modified easily to create other detection reagents that allow
the pump to respond selectively to other analytes as well,[19]
thus providing access to tunable, analyte-selective pumps. To
demonstrate this capability, we exposed the TBS-PPHA
pump to a 75 mm phosphate-buffered solution (pH 6.8, 1.0%
w/v BSA, 258C) that contained 3.1 mm of reagent 1. In the
absence of b-d-glucuronidase, no pumping was observed, but
in the presence of 3.1mm b-d-glucuronidase (Supporting
Information, Movie S5), the average pumping velocity was
1.3 ꢁ 0.5 mmsꢀ1, which is approximately the maximum esti-
mated pumping velocity for the TBS-PPHA pump (Support-
ing Information, Figure S3 and Movie S5). A control experi-
ment that contained the enzyme but not reagent 1 did not lead
to pumping, further confirming that both reagent 1 and TBS-
PPHA are needed to create an enzyme-responsive pump
(Supporting Information, Movie S5).
In conclusion, we have described a set of single-use,
analyte-responsive microscale pumps. These pumps are
capable of moving fluids and micrometer-scale particles
autonomously with velocities that are dictated by the
concentration of the specific analyte that turns “on” the
pump. We also demonstrated that the pairing of a small-
molecule detection reagent with the responsive pumps offers
an opportunity to vary the selectivity of the pump in response
to a variety of enzymes or chemical signals. These types of
pumps will be useful in diverse applications ranging from
microanalysis and microfluidics, to single-use diagnostic
devices. Experiments to demonstrate these types of applica-
tions are currently in progress.
Figure 5. Configuration of a channel to demonstrate that the TBS-
PPHA pump is capable of moving particles over distances of about
5 mm and around corners. The gradient of color from red to pink
represents an approximation of the concentration of monomer within
the channel.
fluids long distances through channels (Figure 5; Supporting
Information, Movie S4). In this experiment, we evaluated
whether the tracer particles could move from right to left (the
pump was positioned on the right), turn left, and then move
from right to left again, but in a perpendicular direction in the
x,y plane; we also tested whether the particles could be
pumped the entire length of the 5 mm-long channel.
Upon exposure to fluoride, the pumping was confined
initially from right to left in the top channel, as expected, but
over time (that is, presumably once the concentration
gradient of the monomeric products increased), the particles
in the left chamber turned the corner and started to flow in
the second channel, again from right to left, until they reached
the end of the channel (Supporting Information, Movie S4).
The estimated power for this type of pump is roughly 4 ꢀ
10ꢀ18Jsꢀ1 (see the Supporting Information). The ability to
pump over this distance and around corners suggests that
these types of analyte-responsive, self-powered pumps may
be suitable for use in smart devices with complex channel
architectures.
Finally, we demonstrated that the TBS-PPHA pumps can
be made responsive to specific analytes other than fluoride,
which is a critical capability for future applications. For
example, Figure 6 shows a reagent (1) that responds to the
Received: November 4, 2011
Revised: December 19, 2011
Published online: February 2, 2012
Keywords: depolymerization · diffusiophoresis ·
.
microfluidic pumps · molecular devices · sensors
[1] J. Wang, ACS Nano 2009, 3, 4 – 9.
[2] Y. Hong, M. Diaz, U. M. Cꢁrdova-Figueroa, A. Sen, Adv. Funct.
Mater. 2010, 20, 1568 – 1576.
[3] a) M. E. Ibele, Y. Wang, T. R. Kline, T. E. Mallouk, A. Sen, J.
Am. Chem. Soc. 2007, 129, 7762 – 7763; b) A. A. Solovev, S.
Sanchez, Y. Mei, O. G. Schmidt, Phys. Chem. Chem. Phys. 2011,
13, 10131 – 10135; c) T. R. Kline, W. F. Paxton, Y. Wang, D.
Velegol, T. E. Mallouk, A. Sen, J. Am. Chem. Soc. 2005, 127,
17150 – 17151.
[4] a) A. P. Esser-Kahn, N. R. Sottos, S. R. White, J. S. Moore, J. Am.
Chem. Soc. 2010, 132, 10266 – 10268; b) A. P. Esser-Kahn, S. A.
Odom, N. R. Sottos, S. R. White, J. S. Moore, Macromolecules
2011, 44, 5539 – 5553; c) A. Sagi, R. Weinstain, N. Karton, D.
Shabat, J. Am. Chem. Soc. 2008, 130, 5434 – 5435; d) M. A.
DeWit, E. R. Gillies, J. Am. Chem. Soc. 2009, 131, 18327 – 18334.
Figure 6. A pump that turns “on” in response to the enzyme b-d-
glucuronidase. This type of enzyme-responsive pump requires the
TBS-PPHA film and a detection reagent (for example, reagent 1) that
releases fluoride in response to the enzyme.
Angew. Chem. Int. Ed. 2012, 51, 2400 –2404
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 2403

.
Angewandte
Communications
[5] W. Seo, S. T. Phillips, J. Am. Chem. Soc. 2010, 132, 9234 – 9235.
[6] E. M. Purcell, Am. J. Phys. 1977, 45, 3 – 11.
[10] S. A. MacDonald, C. G. Willson, J. M. J. Frechet, Acc. Chem.
Res. 1994, 27, 151 – 158.
[11] a) D. C. Smith, J. Biomed. Mater. Res. 1971, 5, 189 – 205; b) J.
Kreuter, S. N. Mills, S. S. Davis, C. G. Wilson, Int. J. Pharm. 1983,
16, 105 – 113.
[12] a) F. Leonard, R. K. Kulkarni, G. Brandes, J. Nelson, J. J.
Cameron, J. Appl. Polym. Sci. 1966, 10, 259 – 272; b) C.
Limouzin, A. Caviggia, F. Ganachaud, P. Hꢃmery, Macromole-
cules 2003, 36, 667 – 674.
[13] P. D. Frymier, R. M. Ford, H. C. Berg, P. T. Cummings, Proc.
Natl. Acad. Sci. USA 1995, 92, 6195 – 6199.
[14] R. A. Pavlick, S. Sengupta, T. McFadden, H. Zhang, A. Sen,
Angew. Chem. 2011, 123, 9546 – 9549; Angew. Chem. Int. Ed.
2011, 50, 9374 – 9377.
[7] a) J. E. Avron, O. Kenneth, D. H. Oaknin, New J. Phys. 2005, 7,
234; b) C. M. Pooley, G. P. Alexander, J. M. Yeomans, Phys. Rev.
Lett. 2007, 99, 228103; c) R. Golestanian, A. Ajdari, Phys. Rev.
Lett. 2008, 100, 038101; d) T. R. Kline, J. Iwata, P. E. Lammert,
T. E. Mallouk, A. Sen, D. Velegol, J. Phys. Chem. B 2006, 110,
24513 – 24521; e) W. F. Paxton, K. C. Kistler, C. C. Olmeda, A.
Sen, S. K. St. Angelo, Y. Cao, T. E. Mallouk, P. E. Lammert,
V. H. Crespi, J. Am. Chem. Soc. 2004, 126, 13424 – 13431; f) W. F.
Paxton, A. Sen, T. E. Mallouk, Chem. Eur. J. 2005, 11, 6462 –
6470; g) W. F. Paxton, P. T. Baker, T. R. Kline, Y. Wang, T. E.
Mallouk, A. Sen, J. Am. Chem. Soc. 2006, 128, 14881 – 14888;
h) Y. Hong, D. Velegol, N. Chaturvedi, A. Sen, Phys. Chem.
Chem. Phys. 2010, 12, 1423 – 1435; i) T. Mirkovic, N. S. Zacharia,
G. D. Scholes, G. A. Ozin, Small 2010, 6, 159 – 167.
[15] P. O. Staffeld, J. A. Quinn, J. Colloid Interface Sci. 1989, 130, 88 –
100.
[16] J. L. Anderson, Annu. Rev. Fluid Mech. 1989, 21, 61 – 99.
[17] J. R. Howse, R. A. L. Jones, A. J. Ryan, T. Gough, R. Vafa-
bakhsh, R. Golestanian, Phys. Rev. Lett. 2007, 99, 048102.
[18] L. Fiksdal, I. Tryland, Curr. Opin. Biotechnol. 2008, 19, 289 – 294.
[19] M. S. Baker, S. T. Phillips, J. Am. Chem. Soc. 2011, 133, 5170 –
5173.
[8] a) V. Lenaerts, P. Couvreur, D. Christiaens-Leyh, E. Joiris, M.
Roland, B. Rollman, P. Speiser, Biomaterials 1984, 5, 65 – 68;
b) R. H. Mꢂller, C. Lherm, J. Herbert, P. Couvreur, Biomaterials
1990, 11, 590 – 595.
[9] H. Ito, C. G. Willson, Polym. Eng. Sci. 1983, 23, 1012 – 1018.
2404 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 2400 –2404